Durable left ventricular assist device support as a bridge to heart transplant candidacy†

Abstract

Open in new tabDownload slide

OBJECTIVES

Left ventricular assist devices are funded in the UK exclusively as a bridge to transplant (BTT). However, patients who potentially could receive a transplant may develop reversible contraindications to transplant. Bridge to candidacy (BTC) has sometimes been controversial, given the uncertain clinical efficacy of BTC and the risk that reimbursement could be denied. We analysed the UK ventricular assist device database to understand how common BTC was and to assess patient survival rates and incidences of transplants.

METHODS

We identified BTC implants in patients with pulmonary hypertension, chronic kidney disease and obesity using the UK guidelines for heart transplants.

RESULTS

A total of 306 of 540 patients had complete data and 157 were identified as BTC (51%). Overall, there was no difference in survival rates between patients designated as BTC and those designated at BTT (71.9 vs 72.9% at 1 year, respectively; P = 0.82). However, the survival rate was lower at all time points in those with an estimated glomerular filtration rate (eGFR) <40 and in patients with a body mass index (BMI) >32 up to 1-year postimplant. There were no significant differences in the incidence of transplant between patients who were BTC and BTT or for any subgroup up to 5 years. However, we noted a diverging trend towards a lower cumulative incidence of transplant for patients with a BMI >32.

CONCLUSIONS

BTC is common in the UK and appears clinically effective, given that the survival rates and the incidence of transplants were comparable with those for BTT. Patients with a high BMI have a worse survival rate through to 1 year and a trend for a lower incidence of a transplant. Patients with a low eGFR also have a worse survival rate, but a similar proportion received transplants.

INTRODUCTION

Durable implantable left ventricular assist devices (LVADs) have been shown to be superior to medical therapy for the treatment of advanced chronic heart failure [1], and survival improved substantially following the introduction of continuous flow devices [2]. However, long-term survival remains inferior to that after a heart transplant [3, 4], and there are concerns about the affordability and cost-effectiveness of this therapy [5]. Therefore, although they are supported by current heart failure guidelines (e.g. European [6]), access to durable implantable LVADs are restricted by some publicly funded health care systems, and availability varies considerably from country to country.

One form of restriction is to link the use of an LVAD to a patient’s eligibility for a heart transplant, the bridge-to-transplant (BTT) strategy [7]. This method is currently used by the UK National Health Service (NHS). Unfortunately, in the current era of chronic heart failure therapy, many patients with advanced chronic heart failure who could potentially be eligible for a transplant have already developed complications that are relative or absolute contraindications to a heart transplant before being referred to a transplant centre. Furthermore, owing to the scarcity of suitable donor hearts, other patients deteriorate while waiting for a heart transplant. It has been shown that some comorbidities associated with advanced chronic heart failure may be alleviated by LVAD therapy [the bridge-to-candidacy (BTC) strategy]. However, because these patients do not meet standard heart transplant criteria at the moment when they require LVAD support, there has sometimes been controversy surrounding this strategy and a risk that reimbursement could be denied.

LVADs are reportedly effective in reversing secondary pulmonary hypertension and elevated pulmonary vascular resistance (PVR) [8–10]. These haemodynamic measurements are regarded as a relative contraindication to a heart transplant [owing to the risk of right ventricular (RV) failure], especially when the pulmonary artery systolic pressure is above 60 mmHg and the PVR is above 5 Wood units [11]. In cases of cardiorenal syndrome, where the creatinine clearance is considered too low for transplant yet intrinsic renal disease has been reasonably excluded, LVADs are often successful in improving renal function through the sustained provision of improved cardiac output [12]. It has nevertheless been reported that impaired renal function prior to LVAD implant is associated with worse overall survival [13, 14]. BTC in the context of obesity remains a subject of debate due to observations that obese patients supported with continuous flow LVADs rarely lose weight [15, 16]. Furthermore, although several single-centre observational studies show no impact of obesity on outcomes after LVAD, the INTERMACS registry identifies obesity as a risk factor for increased mortality after implant [17]. Various centres also report increased complication rates with obesity such as driveline infection and rehospitalisation rates, particularly when body mass index (BMI) is >35 [18, 19]. Whether obesity as a qualifying indication for BTC influences the subsequent transplant rate is uncertain.

BTC has not previously been explicitly stated to be a funded indication for LVAD implantation within the NHS, and it is unknown how commonly it is used within the 6 transplant centres of the UK. Rather, each centre must follow the criteria for implant as set out by the UK National Service Specification for Ventricular Assist Devices as a Bridge to Heart Transplantation or Myocardial Recovery. According to this document, patients should be assessed by a multidisciplinary team and must be listed for transplant but are unlikely to survive long enough to receive a heart despite the availability of the national urgent heart allocation scheme.

Using the UK Ventricular Assist Device (VAD) Database, we sought to identify how common the strategy of BTC was and what the subsequent survival rate and transplant incidence were compared with those of patients being treated as BTT.

METHODS

We analysed data from the UK National VAD database. The cohort included adult patients first implanted with a durable LVAD between 1 January 2007 and 31 December 2015.

We defined the strategy to be BTC if one of the common contraindications [as determined by the UK Guidelines for Heart Transplantation [20], or the International Society for Heart and Lung Transplantation (ISHLT) [21]] was identified preoperatively. The criteria used were the presence of 1 or more of the following:

1. Patients with a pulmonary artery systolic pressure >60 mmHg and/or a PVR >5 Wood units.

2. Patients with a preoperative estimated glomerular filtration rate (eGFR) <40 ml/min/1.73 m².

3. Patients with BMI >32 kg/m².

Patients were classified as BTT if none of these criteria were met (destination therapy is not currently covered by the UK National Health Service). Patients were excluded if at least one of these measurements was missing. Categorical variables were compared using the χ² test and continuous variables by the Wilcoxon rank sum. Outcomes on support were analysed using competing risks methods to estimate the cumulative incidence of transplant, explant, death or remaining on support over time. Cumulative incidence curves for incidence of transplant were compared between the BTC and BTT groups using the method of Gray [22]. Additional subgroup comparisons were made between patients who met each BTC criterion, compared with the rest, and patients meeting 1 criterion compared with more than 1. Overall patient survival was then estimated using the Kaplan–Meier method where survival data were not censored for competing events such as transplant or explant for recovery (hence death was counted at any time whether on support, or postexplant, postlisting, or post-transplantation). Kaplan–Meier curves were compared using the log-rank test. The data analysis was carried out using the SAS version 9.4 software (SAS Institute Inc., Cary, NC, USA).

RESULTS

A total of 540 patients were identified as having an LVAD implant during the study period. Of these, 311 patients had complete preimplant data to allow determination of BTC versus BTT, but 5 of these patients did not have sufficient information postimplant and were excluded from the competing risks analysis. Our total study cohort number was therefore 306. The characteristics of the whole group and of the BTC and BTT groups are shown in Table 1. Dilated cardiomyopathy was the most common aetiology; most patients had an ejection fraction below 20% and required continuous intravenous inotropes. We noted that patients in the BTT group were sicker compared to those in the BTC group, with 56% of the former being in INTERMACS 1 or 2 profile compared to 38% in the latter. Intra-aortic balloon pumps were used preoperatively in more BTT patients, which reflected this difference. However, there were more patients with diabetes in the BTC group.

The most commonly used LVAD in this study was the HeartWare HVAD (Medtronic, Minneapolis, MN, USA) in 62%, followed by HeartMate II (Abbott Laboratories, Chicago, IL, USA) in 28%. We noted that the proportion of patients receiving a HeartMate II was higher in the BTC group compared to the BTT group. Overall, 26% died on support at 1-year postimplant and 48% died by 5 years. Very few of the patients were explanted, and the initial transplant incidence was low, but it did reach 35% by 5 years on support. There was no statistically significant difference between BTT and BTC in respect to the incidence of an urgent transplant (χ² test, P = 0.66). The cumulative incidence function for transplant listing was 488 days for the BTC group (95% confidence interval 245–731).

One hundred and fifty-seven (51%) of the patients were classified as BTC. Survival was similar at all time points between the patients in the BTC and BTT groups (Fig. 1 and Table 2). At 30 days, 87.9% were alive in the BTC group compared to 91.3% in the BTT group (P = 0.34). At 1 year, survival was 71.9% for those in the BTC compared to 72.9% for those in the BTT group (P = 0.82). Survival at 2 years was 58.0% and 62.2%, respectively (P = 0.48). The incidence of transplant between the BTC and BTT groups (Fig. 2, Table 3) was also similar at all time points. By 5 years, 33% of BTC patients had received a transplant compared to 38% of BTT patients (P = 0.33). If patients had more than 1 criterion for BTC, their survival on support and incidence of transplant at all time points remained similar to those of BTT patients.

Analysis of the bridge to candidacy subgroups

Survival was significantly worse for patients with a preoperative eGFR <40 ml/min/1.73 m² [median eGFR 33.7, interquartile range (IQR) 27.1–36.6] compared with that for patients with preserved renal function (median eGFR 68.1, IQR 54.5–88.7). This was true at all time points with the respective results for 30 days: 75.6% vs 91.7% (P = 0.001), at 1 year: 58.5% vs 74.5% (P = 0.011) and at 2 years: 47.2% vs 62.0% (P = 0.015). The survival curves for both groups can be seen in Fig. 3 with corresponding data shown in Table 4. However, interestingly, eGFR did not appear to alter the incidence of transplant at any time point (Fig. 4 and Supplementary Material, Table S1). By 5 years, 34% of the group with a low eGFR had received transplants compared to 35% of the group with the higher eGFR (P = 0.79).

We observed reduced survival rates in patients with a BMI >32 compared to those in patients with a BMI ≤32 (Supplementary Material, Fig. S1 and Supplementary Material, Table S2). At 30 days and 1 year, the survival rates were 80.6% vs 90.7% (P = 0.059) and 58.3% vs 74.3% (P = 0.044), respectively. At 2 years, the survival rate was 54.9% vs 60.8%, although at this stage there was no statistical significance (P = 0.29). A diverging trend in the cumulative incidence of transplant was observed in patients in the high BMI group compared with those in the low BMI group. By 5 years, 18% vs 37% of the patients, respectively, had received a transplant, although this result missed statistical significance (P = 0.090).

The group of patients with severe pulmonary hypertension and/or high PVR had a median PA systolic pressure of 67 mmHg (IQR 62–75 mmHg). Their survival compared to patients with lower pulmonary pressure and PVR (median PA systolic pressure 45 mmHg, IQR 36–52 mmHg) at 30 days was 93.7% vs 87.2%, respectively (P = 0.074). At 1 year, the survival rate was 77.3% vs 69.5% (P = 0.12) and at 2 years, 63.2% vs 58.2% (P = 0.28), respectively. No significant differences were identified for the incidence of transplant at any time point.

DISCUSSION

We observed that BTC is a common strategy for the implantation of an LVAD in the UK, representing over half of all LVAD implants. We also noted in the first annual report of the ISHLT Mechanically Assisted Circulatory Support (IMACS) Registry that the incidence of BTC has now surpassed that of BTT as a strategy for implant [4]. These observations underscore an important clinical reality for transplant teams. In the current era of chronic heart failure therapy, many young patients who are considered for transplant have already progressed to a stage in their condition where the ideal time window for a transplant has passed. Advances in medical and pacemaker therapies allow patients to live longer with heart failure and with more stability than was possible in the past. The trajectory of the condition has therefore been modified in a way that could explain such observations. However, we found that the use of LVADs as BTC was clinically effective. In particular, we saw that BTC patients had equivalent survival rates and a similar incidence of transplants compared to BTT patients. This observation is important because the UK NHS has only been willing to cover the costs of LVAD implants when the implanting team declares an intention to transplant the patient whereas many patients with advanced heart failure fail to meet the standard heart transplant guidelines. Nevertheless, different patterns of outcomes could be discerned in the BTC subgroups.

Patients with impaired renal function (defined as eGFR <40 ml/min/1.73 m²) represented 26% of the BTC cohort. This group had worse survival rates at all time points after implantation compared to those with preserved renal function. Impaired renal function is therefore an important risk factor to consider prior to an implant and may stimulate further interest in the concept of preoperative optimization prior to surgery. Many centres consider short-term support or extracorporeal membrane oxygenation for patients in severe renal failure prior to committing to a durable LVAD. However, the exact time for considering a short-term intervention and whether it improves the overall outcomes is a subject of debate outside the remit of our article. Of note, the median eGFR in the impaired group was 33.7, meaning that the majority of patients were not in the advanced stages of renal failure. Similar observations have been made in other studies documenting the impact of renal impairment on survival after receiving an LVAD [13, 14]. Nevertheless, although the overall incidence of transplant for the entire cohort was low (35% at 5 years), we observed that patients with impaired renal function who survive with an LVAD do not appear to be disadvantaged for a future transplant.

Severe pulmonary hypertension and/or elevated PVR were common criteria for BTC implantation. Pulmonary artery systolic pressures in excess of 60 mmHg were present in 63% of the BTC cohort and in 32% of all patients with an LVAD. This group of patients had survival and transplant incidences comparable to those of patients with lower PA pressure and PVR. LVADs are known to be very effective at reversing this contraindication to transplant, meaning most patients can be listed postimplant [9, 10]. In the early postoperative period, we observed a good survival rate (93.7%), although the difference with BTT patients (87.2%) narrowly failed to make statistical significance, so we cannot make firm conclusions. We speculate that elevated pulmonary systolic pressure is usually an indication of preserved RV function and therefore the risk of early RV failure is often low.

Patients with a high BMI (>32 kg/m²) accounted for 23% of the BTC cohort and 11.8% of all LVAD implants. BTC in the context of obesity remains a subject of debate due to observations that the majority of obese patients supported with continuous flow LVADs do not appear to lose a significant amount of weight [15, 16]. Furthermore, although several single-centre observational studies show no impact of obesity on outcomes after LVAD, the IMACS registry identifies obesity as a significant risk factor for increased mortality after LVAD implant [4]. Various studies also report increased driveline infection risk and rehospitalisation rate with obesity, particularly when BMI is >35 [18, 19]. In our study, those with an elevated BMI (median 33.3 kg/m²) had a worse survival rate through to 1 year after implant compared to those with a lower BMI (median 25.3 kg/m²). We also observed a diverging trend towards a lower incidence of transplant at 5 years, although the result at this time did not quite make statistical significance.

This is a retrospective observational study of a national registry, which therefore comes with method limitations, including the potential of selection bias. This limitation must be taken onto account and caution should be applied in attempting to translate the findings to reimbursement decisions in other regions. However, unlike some other VAD registries, the UK VAD database is a mandatory requirement for each centre and is regularly audited by NHS England to check for conformity. Although the criteria for BTC were defined retrospectively, the audit was performed by a clinical audit group containing a representative from each UK centre and was entirely derived from the UK Transplant Guidelines [20] and ISHLT [21]. Yet we acknowledge that the transplant centre’s multidisciplinary team makes the ultimate decision over candidacy. Many of the factors listed as BTC criteria are in fact relative (not absolute) contraindications to transplant.

We also wish to highlight a strength of the analysis of our UK data. Unlike most other VAD registries such as INTERMACS, EUROMACS and IMACS, we are able in the UK to follow survival without censoring for subsequent interventions such as transplant or explant for recovery. Although this therefore might be a contributing factor to the appearance of a relatively low survival rate of 72% at 1 year, it is nevertheless a more accurate description of survival after the implant of an LVAD.

Our analysis includes implants performed up to 31 December 2015, which covers a period before the 2016 ISHLT listing criteria for heart transplants gave modified guidance on BMI and eGFR [21]. It was not possible to identify recent or active cancer diagnoses within the database or an accurate smoking history, so we have no data to present for these groups of patients. The ISHLT currently recommends that patients with a history of cancer should have individualized management with respect to eligibility for a transplant [21]. The previous 5-year remission threshold after identification of a solid organ tumour to safely proceed with a transplant has been described as arbitrary. Instead it is recommended that collaboration with oncology should occur to assess each patient as to his or her risk of tumour recurrence. With regards to LVAD therapy, those who are currently deemed unsuitable for transplant (due to a recent or active cancer diagnosis) could theoretically be considered for a device as BTC, to offer them the chance of proving sustained remission before being listed for a transplant (at present, the NHS does not cover LVAD therapy in this patient group). It remains uncertain what the survival and transplant rates are for these patients. Finally, our analysis was unable to specifically evaluate the incidence of RV failure between the subgroups to determine the reasons behind any given trend of survival advantage.

CONCLUSION

The use of durable LVADs in the setting of potentially reversible contraindications to transplant has been widely adopted. We have observed that the outcomes for this group of patients are comparable to those for the standard BTT group for both survival rate and rate of transplant. However, the efficacy of a BTC strategy in patients with obesity, cancer and a history of smoking remains uncertain.

Funding

The UK VAD database is funded by the NHS England. The NHS Blood and Transplant division houses the database and provides statistical support for the analysis. Interpretation is the responsibility of the authors (i.e. our comments are not NHSE policy).

Conflict of interest: We have declared that Shaw, Venkateswaran and Schueler consult/proctor for Abbott and Medtronic.

Footnotes

REFERENCES